Does changing the anomaly detection setting from auto to manual stop the baseline generation?

[waikeat_chan]

In case in the future when we switch back from manual to auto, we're wondering do we need to wait for another '20% of one week', or can immediately get the baseline back.

[rodrigo_alvare1 · Accepted solution]

It will definitely affect the autodetected baseline and you will need time to train again the autodetection of your baseline. A best practice is to leave the autobaseline as it is and create Custom alerts for your SLAs metrics.

With this approach you will be able to get alert on the conditions you want and still get information of the current behavior or your service (and together with the AI will alert you when something is abnormal)

[waikeat_chan]

Thanks for the answer Rodrigue, this is what I think as well, so now I can reply customer confidently.

Another question though: I did try custom alert, but seems like not able to utilize tags to filter the scope of the custom alert. Do you also face the issue of not able to utilize tagging as filter in creating custom alert?

Best Regards,

Wai Keat

[kyle_kowalski]

The alerting profile that you set up will look at the tags for those custom alerts.